Defining the rhythm: a new method to classify tremor and myoclonus

BackgroundTremor hallmark feature is rhythmicity, which can be quantified using power spectral density analysis. However, tremor exhibits considerable variability, ranging from highly regular to more irregular patterns. Similarly, rhythmicity in myoclonus also varies, but it typically manifesting as arrhythmic jerks. ObjectivesTo develop power spectral density-based measures of movement regularity for the classification tremor and myoclonus. MethodsElectromyography data from 153 patients were analysed retrospectively, including orthostatic tremor (n=36), essential tremor (n=40), dystonic tremor (n=42), and limb cortical myoclonus (n=35). Five power spectral density analysis-derived variables were assessed: peak prominence, peak-to-broadband power ratio, peak frequency, peak width, and harmonics. Discriminant analysis evaluated classification accuracy across groups. ResultsPeak prominence was highest in orthostatic tremor and higher in essential tremor than dystonic tremor or myoclonus. Peak-to-broadband power ratio showed similar trends. Peak frequency differed across groups, with myoclonus highest and orthostatic tremor exceeding essential tremor and dystonic tremor. Peak width was larger in myoclonus and, to a less extent, in dystonic tremor compared to essential tremor. Harmonics were greater in orthostatic tremor and essential tremor compared to dystonic tremor and myoclonus. Discriminant analysis correctly classified 86.3% of cases, with overlap between essential tremor and dystonic tremor. Receiver operating characteristic curve analysis for peak prominence and width demonstrated high classification accuracy between essential tremor and dystonic tremor. ConclusionsOur findings represent a promising initial step toward establishing objective, power spectral density-based measures for the classification of tremor and myoclonus. These tools could enhance diagnostic accuracy and deepen insights into these disorders.